4 Matching Annotations
  1. Jun 2020
  2. Dec 2018
    1. ayesian topic mod-els, the most popular such models today are variants of Latent Dirichlet Allocation (LDA, Bleiet al., 2003b), which provides a way to automatically discover latent or implicittopicsin otherwiseunstructured collections of text
    2. Recasens et al. (2013) draw a related distinction betweenframing bias, which involves explicitlysubjective words or phrases linked with a particular point of view, andepistemological bias, whichinvolves implicit assumptions and presuppositions in ostensibly neutral writing.)
    3. Policy Frames Codebook